Add the possibility to define rewrite rules for each feed

reader/rewrite/rewrite_functions.go

@@ -0,0 +1,40 @@
+// Copyright 2017 Frédéric Guillot. All rights reserved.
+// Use of this source code is governed by the Apache 2.0
+// license that can be found in the LICENSE file.
+
+package rewrite
+
+import (
+	"regexp"
+	"strings"
+
+	"github.com/PuerkitoBio/goquery"
+)
+
+var (
+	youtubeRegex = regexp.MustCompile(`youtube\.com/watch\?v=(.*)`)
+)
+
+func addImageTitle(entryURL, entryContent string) string {
+	doc, err := goquery.NewDocumentFromReader(strings.NewReader(entryContent))
+	if err != nil {
+		return entryContent
+	}
+
+	imgTag := doc.Find("img").First()
+	if titleAttr, found := imgTag.Attr("title"); found {
+		return entryContent + `<blockquote cite="` + entryURL + `">` + titleAttr + "</blockquote>"
+	}
+
+	return entryContent
+}
+
+func addYoutubeVideo(entryURL, entryContent string) string {
+	matches := youtubeRegex.FindStringSubmatch(entryURL)
+
+	if len(matches) == 2 {
+		video := `<iframe width="650" height="350" frameborder="0" src="https://www.youtube-nocookie.com/embed/` + matches[1] + `" allowfullscreen></iframe>`
+		return video + "<p>" + entryContent + "</p>"
+	}
+	return entryContent
+}

reader/rewrite/rewriter.go

@@ -5,44 +5,39 @@
 package rewrite
 
 import (
-	"regexp"
 	"strings"
 
-	"github.com/PuerkitoBio/goquery"
+	"github.com/miniflux/miniflux2/url"
 )
 
-var rewriteRules = []func(string, string) string{
-	func(url, content string) string {
-		re := regexp.MustCompile(`youtube\.com/watch\?v=(.*)`)
-		matches := re.FindStringSubmatch(url)
-
-		if len(matches) == 2 {
-			video := `<iframe width="650" height="350" frameborder="0" src="https://www.youtube-nocookie.com/embed/` + matches[1] + `" allowfullscreen></iframe>`
-			return video + "<p>" + content + "</p>"
-		}
-		return content
-	},
-	func(url, content string) string {
-		if strings.HasPrefix(url, "https://xkcd.com") {
-			doc, err := goquery.NewDocumentFromReader(strings.NewReader(content))
-			if err != nil {
-				return content
-			}
-
-			imgTag := doc.Find("img").First()
-			if titleAttr, found := imgTag.Attr("title"); found {
-				return content + `<blockquote cite="` + url + `">` + titleAttr + "</blockquote>"
-			}
-		}
-		return content
-	},
-}
-
 // Rewriter modify item contents with a set of rewriting rules.
-func Rewriter(url, content string) string {
-	for _, rewriteRule := range rewriteRules {
-		content = rewriteRule(url, content)
+func Rewriter(entryURL, entryContent, customRewriteRules string) string {
+	rulesList := getPredefinedRewriteRules(entryURL)
+	if customRewriteRules != "" {
+		rulesList = customRewriteRules
 	}
 
-	return content
+	rules := strings.Split(rulesList, ",")
+	for _, rule := range rules {
+		switch strings.TrimSpace(rule) {
+		case "add_image_title":
+			entryContent = addImageTitle(entryURL, entryContent)
+		case "add_youtube_video":
+			entryContent = addYoutubeVideo(entryURL, entryContent)
+		}
+	}
+
+	return entryContent
+}
+
+func getPredefinedRewriteRules(entryURL string) string {
+	urlDomain := url.Domain(entryURL)
+
+	for domain, rules := range predefinedRules {
+		if strings.Contains(urlDomain, domain) {
+			return rules
+		}
+	}
+
+	return ""
 }

reader/rewrite/rewriter_test.go

@@ -7,7 +7,7 @@ package rewrite
 import "testing"
 
 func TestRewriteWithNoMatchingRule(t *testing.T) {
-	output := Rewriter("https://example.org/article", `Some text.`)
+	output := Rewriter("https://example.org/article", `Some text.`, ``)
 	expected := `Some text.`
 
 	if expected != output {
@@ -16,7 +16,7 @@ func TestRewriteWithNoMatchingRule(t *testing.T) {
 }
 
 func TestRewriteWithYoutubeLink(t *testing.T) {
-	output := Rewriter("https://www.youtube.com/watch?v=1234", `Video Description`)
+	output := Rewriter("https://www.youtube.com/watch?v=1234", `Video Description`, ``)
 	expected := `<iframe width="650" height="350" frameborder="0" src="https://www.youtube-nocookie.com/embed/1234" allowfullscreen></iframe><p>Video Description</p>`
 
 	if expected != output {
@@ -24,11 +24,37 @@ func TestRewriteWithYoutubeLink(t *testing.T) {
 	}
 }
 
+func TestRewriteWithInexistingCustomRule(t *testing.T) {
+	output := Rewriter("https://www.youtube.com/watch?v=1234", `Video Description`, `some rule`)
+	expected := `Video Description`
+	if expected != output {
+		t.Errorf(`Not expected output: got "%s" instead of "%s"`, output, expected)
+	}
+}
+
 func TestRewriteWithXkcdLink(t *testing.T) {
 	description := `<img src="https://imgs.xkcd.com/comics/thermostat.png" title="Your problem is so terrible, I worry that, if I help you, I risk drawing the attention of whatever god of technology inflicted it on you." alt="Your problem is so terrible, I worry that, if I help you, I risk drawing the attention of whatever god of technology inflicted it on you." />`
-	output := Rewriter("https://xkcd.com/1912/", description)
+	output := Rewriter("https://xkcd.com/1912/", description, ``)
 	expected := description + `<blockquote cite="https://xkcd.com/1912/">Your problem is so terrible, I worry that, if I help you, I risk drawing the attention of whatever god of technology inflicted it on you.</blockquote>`
 	if expected != output {
 		t.Errorf(`Not expected output: got "%s" instead of "%s"`, output, expected)
 	}
 }
+func TestRewriteWithXkcdLinkAndNoImage(t *testing.T) {
+	description := "test"
+	output := Rewriter("https://xkcd.com/1912/", description, ``)
+	expected := description
+	if expected != output {
+		t.Errorf(`Not expected output: got "%s" instead of "%s"`, output, expected)
+	}
+}
+
+func TestRewriteWithXkcdAndNoImage(t *testing.T) {
+	description := "test"
+	output := Rewriter("https://xkcd.com/1912/", description, ``)
+	expected := description
+
+	if expected != output {
+		t.Errorf(`Not expected output: got "%s" instead of "%s"`, output, expected)
+	}
+}

reader/rewrite/rules.go

@@ -0,0 +1,30 @@
+// Copyright 2017 Frédéric Guillot. All rights reserved.
+// Use of this source code is governed by the Apache 2.0
+// license that can be found in the LICENSE file.
+
+package rewrite
+
+// List of predefined rewrite rules (alphabetically sorted)
+// Available rules: "add_image_title", "add_youtube_video"
+// domain => rule name
+var predefinedRules = map[string]string{
+	"abstrusegoose.com":      "add_image_title",
+	"amazingsuperpowers.com": "add_image_title",
+	"cowbirdsinlove.com":     "add_image_title",
+	"drawingboardcomic.com":  "add_image_title",
+	"exocomics.com":          "add_image_title",
+	"happletea.com":          "add_image_title",
+	"imogenquest.net":        "add_image_title",
+	"lukesurl.com":           "add_image_title",
+	"mercworks.net":          "add_image_title",
+	"mrlovenstein.com":       "add_image_title",
+	"nedroid.com":            "add_image_title",
+	"oglaf.com":              "add_image_title",
+	"optipess.com":           "add_image_title",
+	"peebleslab.com":         "add_image_title",
+	"sentfromthemoon.com":    "add_image_title",
+	"thedoghousediaries.com": "add_image_title",
+	"treelobsters.com":       "add_image_title",
+	"youtube.com":            "add_youtube_video",
+	"xkcd.com":               "add_image_title",
+}